- Pacman-dev - lists.archlinux.org

Re: [pacman-dev] [arch-projects] [mkinitcpio][PATCH 21/26] lsinitcpio: disable color when stdout isn't a tty
by Dan McGee 27 Sep '11

27 Sep '11

On Mon, Sep 26, 2011 at 9:01 PM, Dave Reisner <d(a)falconindy.com> wrote: > On Mon, Sep 26, 2011 at 08:57:40PM -0500, Dan McGee wrote: >> On Mon, Sep 26, 2011 at 8:22 PM, Dave Reisner <d(a)falconindy.com> wrote: >> > Signed-off-by: Dave Reisner <dreisner(a)archlinux.org> >> Why were we using stderr before? Did something else change? > > You really wanna know? Copy/paste job from makepkg. We should fix this > over in makepkg as well. I'm not sure why it's used there either. makepkg -g >> PKGBUILD. Now whether that is a good reason is up for some debate, but you asked. :) -Dan

1 0

[pacman-dev] [PATCH] Aviod comparison between signed and unsigned types
by Allan McRae 26 Sep '11

26 Sep '11

Signed-off-by: Allan McRae <allan(a)archlinux.org> --- src/pacman/pacman.c | 3 ++- 1 files changed, 2 insertions(+), 1 deletions(-) diff --git a/src/pacman/pacman.c b/src/pacman/pacman.c index 40aa444..326664d 100644 --- a/src/pacman/pacman.c +++ b/src/pacman/pacman.c @@ -760,7 +760,8 @@ static void cl_to_log(int argc, char* argv[]) */ int main(int argc, char *argv[]) { - int i, ret = 0; + int ret = 0; + size_t i; struct sigaction new_action, old_action; const int signals[] = { SIGHUP, SIGINT, SIGTERM, SIGSEGV }; #if defined(HAVE_GETEUID) && !defined(CYGWIN) -- 1.7.6.4

2 1

[pacman-dev] [GIT] The official pacman repository branch, master, updated. v4.0.0rc2-12-g80b0f27
by dan＠archlinux.org 26 Sep '11

26 Sep '11

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "The official pacman repository". The branch, master has been updated via 80b0f271254f965db4d94d6bb165c64ef3553588 (commit) via d1fc3aec4c7f7bab30f8ad7dabd7832c7c8570e6 (commit) via 507b01b912d77f0cc5422e689623f1b19c208525 (commit) via b6ccae2d182899eaf199e868963646b3a14f7e8b (commit) via dd82b8d09f554d709f92fcdc7f873cb551b2d205 (commit) via 6adf502f0d0e684c67d7498057083958becf8b19 (commit) via d26d50e664665905d79a3d5fa589b2ca0bec4ba0 (commit) from 2517ba3303489ed35512cdaf2493ec81efd145e7 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- commit 80b0f271254f965db4d94d6bb165c64ef3553588 Author: Dan McGee <dan(a)archlinux.org> Date: Mon Sep 26 17:31:34 2011 -0500 Ensure fileconflict value is actually a string When we switched to a file object and not just a simple string, we missed an update along the way here in target-target conflicts. This patch looks large, but it really comes down to one errant (char *) cast before that has been reworked to explicitly point to the alpm_file_t object. The rest is simply code cleanup. Signed-off-by: Dan McGee <dan(a)archlinux.org> commit d1fc3aec4c7f7bab30f8ad7dabd7832c7c8570e6 Author: Dave Reisner <dreisner(a)archlinux.org> Date: Sun Sep 25 14:32:15 2011 -0400 bash_completion: add completion for pacman-key Signed-off-by: Dave Reisner <dreisner(a)archlinux.org> Signed-off-by: Dan McGee <dan(a)archlinux.org> commit 507b01b912d77f0cc5422e689623f1b19c208525 Author: Dave Reisner <dreisner(a)archlinux.org> Date: Sun Sep 25 15:24:53 2011 -0400 pacman-key: allow get_from to return a default value Keep the non-zero return val to let the caller know that the key wasn't found. Signed-off-by: Dave Reisner <dreisner(a)archlinux.org> Signed-off-by: Dan McGee <dan(a)archlinux.org> commit b6ccae2d182899eaf199e868963646b3a14f7e8b Author: Dave Reisner <dreisner(a)archlinux.org> Date: Sun Sep 25 14:34:36 2011 -0400 pacman-key: simplify writing to config file Signed-off-by: Dave Reisner <dreisner(a)archlinux.org> Signed-off-by: Dan McGee <dan(a)archlinux.org> commit dd82b8d09f554d709f92fcdc7f873cb551b2d205 Author: Allan McRae <allan(a)archlinux.org> Date: Sat Sep 24 15:08:21 2011 +1000 makepkg: normalize whitespace for optdepends in .PKGINFO Many PKGBUILDs use formatting whitespace when specifying optdepends. This is removed when adding a package to a repo-database so the output of "pacman -Si <package>" and "pacman -Qip <package file>" becomes inconsistent. Instead, do the adjustment when creating the .PKGINFO file. Signed-off-by: Allan McRae <allan(a)archlinux.org> Signed-off-by: Dan McGee <dan(a)archlinux.org> commit 6adf502f0d0e684c67d7498057083958becf8b19 Author: Allan McRae <allan(a)archlinux.org> Date: Sat Sep 24 14:53:20 2011 +1000 makepkg: allow versioned optdepends with epoch Allow the specification of versioned optdepends with an epoch. This also (partially) enforces a whitespace between ":" and the description which is required for the future optdepends parsing code. Signed-off-by: Allan McRae <allan(a)archlinux.org> Signed-off-by: Dan McGee <dan(a)archlinux.org> commit d26d50e664665905d79a3d5fa589b2ca0bec4ba0 Author: Dan McGee <dan(a)archlinux.org> Date: Mon Sep 26 15:22:13 2011 -0500 Revamp pacman signal handler * All errors now go to stderr, so do the same here and simplify the writing of the error message. * Add SIGHUP to the handled signal list, and don't repeat code. * Attempt to release the transaction (e.g. remove the lock file) for all of HUP, INT, and TERM. Signals HUP and INT respects transaction state, TERM will immediately terminate the process. Signed-off-by: Dan McGee <dan(a)archlinux.org> ----------------------------------------------------------------------- Summary of changes: contrib/bash_completion.in | 16 +++++++++++++++ lib/libalpm/conflict.c | 30 +++++++++++++--------------- scripts/makepkg.sh.in | 6 ++-- scripts/pacman-key.sh.in | 14 ++++++------ src/pacman/pacman.c | 46 ++++++++++++++++++++++---------------------- 5 files changed, 63 insertions(+), 49 deletions(-) hooks/post-receive -- The official pacman repository

1 0

[pacman-dev] [PATCH] Ensure fileconflict value is actually a string
by Dan McGee 26 Sep '11

26 Sep '11

When we switched to a file object and not just a simple string, we missed an update along the way here in target-target conflicts. This patch looks large, but it really comes down to one errant (char *) cast before that has been reworked to explicitly point to the alpm_file_t object. The rest is simply code cleanup. Signed-off-by: Dan McGee <dan(a)archlinux.org> --- Should we take this opportunity to rework the file conflict struct type to be well typed instead of using strings everywhere? struct new_fileconflict_t { alpm_fileconflict_t *type; alpm_file_t *file; (frontend would be responsible for root path prepend) alpm_pkg_t *target; alpm_pkg_t *ctarget; }; could change package1 and package2 in _alpm_conflict_t also to be actual packages as well instead of strings. Thoughts? -Dan lib/libalpm/conflict.c | 30 ++++++++++++++---------------- 1 files changed, 14 insertions(+), 16 deletions(-) diff --git a/lib/libalpm/conflict.c b/lib/libalpm/conflict.c index 1c02f18..14c23f4 100644 --- a/lib/libalpm/conflict.c +++ b/lib/libalpm/conflict.c @@ -276,30 +276,29 @@ static alpm_list_t *filelist_operation(alpm_filelist_t *filesA, return ret; } -/* Adds alpm_fileconflict_t to a conflicts list. Pass the conflicts list, type - * (either ALPM_FILECONFLICT_TARGET or ALPM_FILECONFLICT_FILESYSTEM), a file - * string, and either two package names or one package name and NULL. This is - * a wrapper for former functionality that was done inline. +/* Adds alpm_fileconflict_t to a conflicts list. Pass the conflicts list, the + * conflicting file path, and either two packages or one package and NULL. */ static alpm_list_t *add_fileconflict(alpm_handle_t *handle, - alpm_list_t *conflicts, alpm_fileconflicttype_t type, const char *filestr, - const char *name1, const char *name2) + alpm_list_t *conflicts, const char *filestr, + alpm_pkg_t *pkg1, alpm_pkg_t *pkg2) { alpm_fileconflict_t *conflict; MALLOC(conflict, sizeof(alpm_fileconflict_t), goto error); - conflict->type = type; - STRDUP(conflict->target, name1, goto error); + STRDUP(conflict->target, pkg1->name, goto error); STRDUP(conflict->file, filestr, goto error); - if(name2) { - STRDUP(conflict->ctarget, name2, goto error); + if(pkg2) { + conflict->type = ALPM_FILECONFLICT_TARGET; + STRDUP(conflict->ctarget, pkg2->name, goto error); } else { + conflict->type = ALPM_FILECONFLICT_FILESYSTEM; STRDUP(conflict->ctarget, "", goto error); } conflicts = alpm_list_add(conflicts, conflict); _alpm_log(handle, ALPM_LOG_DEBUG, "found file conflict %s, packages %s and %s\n", - filestr, name1, name2 ? name2 : "(filesystem)"); + filestr, pkg1->name, pkg2 ? pkg2->name : "(filesystem)"); return conflicts; @@ -416,9 +415,9 @@ alpm_list_t *_alpm_db_find_fileconflicts(alpm_handle_t *handle, alpm_list_t *k; char path[PATH_MAX]; for(k = common_files; k; k = k->next) { - snprintf(path, PATH_MAX, "%s%s", handle->root, (char *)k->data); - conflicts = add_fileconflict(handle, conflicts, - ALPM_FILECONFLICT_TARGET, path, p1->name, p2->name); + alpm_file_t *file = k->data; + snprintf(path, PATH_MAX, "%s%s", handle->root, file->name); + conflicts = add_fileconflict(handle, conflicts, path, p1, p2); if(handle->pm_errno == ALPM_ERR_MEMORY) { FREELIST(conflicts); FREELIST(common_files); @@ -567,8 +566,7 @@ alpm_list_t *_alpm_db_find_fileconflicts(alpm_handle_t *handle, } if(!resolved_conflict) { - conflicts = add_fileconflict(handle, conflicts, - ALPM_FILECONFLICT_FILESYSTEM, path, p1->name, NULL); + conflicts = add_fileconflict(handle, conflicts, path, p1, NULL); if(handle->pm_errno == ALPM_ERR_MEMORY) { FREELIST(conflicts); if(dbpkg) { -- 1.7.6.4

1 1

[pacman-dev] [PATCH] Revamp pacman signal handler
by Dan McGee 26 Sep '11

26 Sep '11

* All errors now go to stdout, so do the same here and simplify the writing of the error message. * Add SIGHUP to the handled signal list, and don't repeat code. * Attempt to release the transaction (e.g. remove the lock file) for all of HUP, INT, and TERM. Signals HUP and INT respects transaction state, TERM will immediately terminate the process. Signed-off-by: Dan McGee <dan(a)archlinux.org> --- src/pacman/pacman.c | 46 +++++++++++++++++++++++----------------------- 1 files changed, 23 insertions(+), 23 deletions(-) diff --git a/src/pacman/pacman.c b/src/pacman/pacman.c index 061d593..40aa444 100644 --- a/src/pacman/pacman.c +++ b/src/pacman/pacman.c @@ -299,26 +299,29 @@ static void handler(int signum) { int out = fileno(stdout); int err = fileno(stderr); + const char *msg; if(signum == SIGSEGV) { - const char *msg1 = "error: segmentation fault\n"; - const char *msg2 = "Internal pacman error: Segmentation fault.\n" + msg = "\nerror: segmentation fault\n" "Please submit a full bug report with --debug if appropriate.\n"; - /* write a error message to out, the rest to err */ - xwrite(out, msg1, strlen(msg1)); - xwrite(err, msg2, strlen(msg2)); + xwrite(err, msg, strlen(msg)); exit(signum); - } else if(signum == SIGINT) { - const char *msg = "\nInterrupt signal received\n"; + } else if(signum == SIGINT || signum == SIGHUP) { + if(signum == SIGINT) { + msg = "\nInterrupt signal received\n"; + } else { + msg = "\nHangup signal received\n"; + } xwrite(err, msg, strlen(msg)); if(alpm_trans_interrupt(config->handle) == 0) { /* a transaction is being interrupted, don't exit pacman yet. */ return; } - /* no commiting transaction, we can release it now and then exit pacman */ - alpm_trans_release(config->handle); - /* output a newline to be sure we clear any line we may be on */ - xwrite(out, "\n", 1); } + /* SIGINT: no commiting transaction, release it now and then exit pacman + * SIGHUP, SIGTERM: release no matter what */ + alpm_trans_release(config->handle); + /* output a newline to be sure we clear any line we may be on */ + xwrite(out, "\n", 1); cleanup(128 + signum); } @@ -757,8 +760,9 @@ static void cl_to_log(int argc, char* argv[]) */ int main(int argc, char *argv[]) { - int ret = 0; + int i, ret = 0; struct sigaction new_action, old_action; + const int signals[] = { SIGHUP, SIGINT, SIGTERM, SIGSEGV }; #if defined(HAVE_GETEUID) && !defined(CYGWIN) /* geteuid undefined in CYGWIN */ uid_t myuid = geteuid(); @@ -775,17 +779,13 @@ int main(int argc, char *argv[]) sigemptyset(&new_action.sa_mask); new_action.sa_flags = 0; - sigaction(SIGINT, NULL, &old_action); - if(old_action.sa_handler != SIG_IGN) { - sigaction(SIGINT, &new_action, NULL); - } - sigaction(SIGTERM, NULL, &old_action); - if(old_action.sa_handler != SIG_IGN) { - sigaction(SIGTERM, &new_action, NULL); - } - sigaction(SIGSEGV, NULL, &old_action); - if(old_action.sa_handler != SIG_IGN) { - sigaction(SIGSEGV, &new_action, NULL); + /* assign our handler to any signals we care about */ + for(i = 0; i < sizeof(signals) / sizeof(signals[0]); i++) { + int signal = signals[i]; + sigaction(signal, NULL, &old_action); + if(old_action.sa_handler != SIG_IGN) { + sigaction(signal, &new_action, NULL); + } } /* i18n init */ -- 1.7.6.3

3 2

[pacman-dev] [PATCH] makepkg: normalize whitespace for optdepends in .PKGINFO
by Allan McRae 24 Sep '11

24 Sep '11

Many PKGBUILDs use formatting whitespace when specifying optdepends. This is removed when adding a package to a repo-database so the output of "pacman -Si <package>" and "pacman -Qip <package file>" becomes inconsistent. Instead, do the adjustment when creating the .PKGINFO file. Signed-off-by: Allan McRae <allan(a)archlinux.org> --- scripts/makepkg.sh.in | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) diff --git a/scripts/makepkg.sh.in b/scripts/makepkg.sh.in index a0b0375..2703177 100644 --- a/scripts/makepkg.sh.in +++ b/scripts/makepkg.sh.in @@ -1134,7 +1134,7 @@ write_pkginfo() { [[ $license ]] && printf "license = %s\n" "${license[@]}" [[ $replaces ]] && printf "replaces = %s\n" "${replaces[@]}" [[ $groups ]] && printf "group = %s\n" "${groups[@]}" - [[ $optdepends ]] && printf "optdepend = %s\n" "${optdepends[@]}" + [[ $optdepends ]] && printf "optdepend = %s\n" "${optdepends[@]//+([[:space:]])/ }" [[ $conflicts ]] && printf "conflict = %s\n" "${conflicts[@]}" [[ $backup ]] && printf "backup = %s\n" "${backup[@]}" -- 1.7.6.3

1 0

[pacman-dev] [PATCH] makepkg: allow versioned optdepends with epoch
by Allan McRae 24 Sep '11

24 Sep '11

Allow the specification of versioned optdepends with an epoch. This also (partially) enforces a whitespace between ":" and the description which is required for the future optdepends parsing code. Signed-off-by: Allan McRae <allan(a)archlinux.org> --- There are only four packages in the Arch repos that do not use a ":" followed by some whitespace before the description in their optdepends. This catches 3/4 of them, with the following is incorrectly seen as valid: optdepends=('pyglet:plotting') It is probably possible to separate a single word description from a version with epoch with a much more complicated regex... scripts/makepkg.sh.in | 4 ++-- 1 files changed, 2 insertions(+), 2 deletions(-) diff --git a/scripts/makepkg.sh.in b/scripts/makepkg.sh.in index 6304a25..a0b0375 100644 --- a/scripts/makepkg.sh.in +++ b/scripts/makepkg.sh.in @@ -1529,8 +1529,8 @@ check_sanity() { eval $(awk '/^[[:space:]]*optdepends=$/,/$[[:space:]]*(#.*)?$/' "$BUILDFILE" | \ sed -e "s/optdepends=/optdepends_list+=/" -e "s/#.*//" -e 's/\\$//') for i in "${optdepends_list[@]}"; do - local pkg=${i%%:*} - if [[ $pkg != +([[:alnum:]><=.+_-]) ]]; then + local pkg=${i%%:[[:space:]]*} + if [[ $pkg != +([[:alnum:]><=.+_-:]) ]]; then error "$(gettext "Invalid syntax for %s : '%s'")" "optdepend" "$i" ret=1 fi -- 1.7.6.3

1 0

[pacman-dev] [GIT] The official pacman repository branch, master, updated. v4.0.0rc2-5-g2517ba3
by dan＠archlinux.org 22 Sep '11

22 Sep '11

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "The official pacman repository". The branch, master has been updated via 2517ba3303489ed35512cdaf2493ec81efd145e7 (commit) via c406949226fe5759473f13b5eb1fcf48c32d67ea (commit) via 1e0ed133f471c35fe6979612a4cb57926223c9b9 (commit) via 907e8af5be7579129a257b4c9952e86a22df9bb9 (commit) via f1beb050a305f4e363d09579359957097c5baed5 (commit) from 43787d00678e00adee54c9ab4b8909d26cba3940 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- commit 2517ba3303489ed35512cdaf2493ec81efd145e7 Author: Dan McGee <dan(a)archlinux.org> Date: Thu Sep 22 16:29:03 2011 -0500 Update Doxyfile and fix some documentation errors caught by Doxygen A few parameters were outdated or wrongly named, and a few things were explicitly linked that Doxygen wasn't able to resolve. Signed-off-by: Dan McGee <dan(a)archlinux.org> commit c406949226fe5759473f13b5eb1fcf48c32d67ea Author: Dan McGee <dan(a)archlinux.org> Date: Thu Sep 22 15:59:18 2011 -0500 Ensure database validity flags are set in invalid signature case We returned the right error code but never set the flags accordingly. Also, now that we can bail early, ensure we set the error code. Signed-off-by: Dan McGee <dan(a)archlinux.org> commit 1e0ed133f471c35fe6979612a4cb57926223c9b9 Author: Dan McGee <dan(a)archlinux.org> Date: Thu Sep 22 15:47:57 2011 -0500 Handle key import errors correctly and with good error messages This adds calls to gpgme_op_import_result() which we were not looking at before to ensure the key was actually imported. Additionally, we do some preemptive checks to ensure the keyring is even writable if we are going to prompt the user to add things to it. Signed-off-by: Dan McGee <dan(a)archlinux.org> commit 907e8af5be7579129a257b4c9952e86a22df9bb9 Author: Dan McGee <dan(a)archlinux.org> Date: Thu Sep 22 14:40:51 2011 -0500 signing: be consistent with returning -1 for error cases This also fixes a segfault found by dave when key_search is unsuccessful; the key_search return code documentation has also been updated to reflect reality. Signed-off-by: Dan McGee <dan(a)archlinux.org> commit f1beb050a305f4e363d09579359957097c5baed5 Author: Dan McGee <dan(a)archlinux.org> Date: Thu Sep 22 12:08:35 2011 -0500 Update configure.ac version to rc2 Signed-off-by: Dan McGee <dan(a)archlinux.org> ----------------------------------------------------------------------- Summary of changes: configure.ac | 2 +- doc/Doxyfile | 143 +++++++++++++++++++++++++++++++--------------- lib/libalpm/alpm.c | 2 +- lib/libalpm/alpm.h | 13 +++-- lib/libalpm/alpm_list.c | 1 - lib/libalpm/be_sync.c | 3 + lib/libalpm/deps.c | 1 + lib/libalpm/signing.c | 73 ++++++++++++++++++------ lib/libalpm/sync.c | 2 +- 9 files changed, 165 insertions(+), 75 deletions(-) hooks/post-receive -- The official pacman repository

1 0

[pacman-dev] [PATCH] signing: fix segfault on unknown key
by Dave Reisner 22 Sep '11

22 Sep '11

Jump to the error label here, the same as other cases when we don't have valid key data to work with. Signed-off-by: Dave Reisner <dreisner(a)archlinux.org> --- lib/libalpm/signing.c | 1 + 1 files changed, 1 insertions(+), 0 deletions(-) diff --git a/lib/libalpm/signing.c b/lib/libalpm/signing.c index 66607fc..2c26f4e 100644 --- a/lib/libalpm/signing.c +++ b/lib/libalpm/signing.c @@ -247,6 +247,7 @@ static int key_search(alpm_handle_t *handle, const char *fpr, err = gpgme_get_key(ctx, fpr, &key, 0); if(gpg_err_code(err) == GPG_ERR_EOF) { _alpm_log(handle, ALPM_LOG_DEBUG, "key lookup failed, unknown key\n"); + CHECK_ERR(); } else if(gpg_err_code(err) != GPG_ERR_NO_ERROR) { _alpm_log(handle, ALPM_LOG_DEBUG, "gpg error: %s\n", gpgme_strerror(err)); -- 1.7.6.3

1 0

[pacman-dev] [GIT] The official pacman repository annotated tag, v4.0.0rc2, created. v4.0.0rc2
by dan＠archlinux.org 22 Sep '11

22 Sep '11

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "The official pacman repository". The annotated tag, v4.0.0rc2 has been created at 58b4f02d2e0f46c6694f701998e549c6cf17f5ed (tag) tagging 43787d00678e00adee54c9ab4b8909d26cba3940 (commit) replaces v4.0.0rc1 tagged by Dan McGee on Thu Sep 22 11:51:26 2011 -0500 - Log ----------------------------------------------------------------- 4.0.0 RC2- import keys, fully test signing -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.11 (GNU/Linux) iEYEABECAAYFAk57Z58ACgkQXC5GoPU6du2y0wCgooXA6y4dtDn0uQN3FGsed2np xLoAoLq8Mr5CViBLqnIj4RQBDKHMM1UU =fKRN -----END PGP SIGNATURE----- Allan McRae (12): repo-add: indicate whether package signature is found repo-add: fix creation of signature symlink configure.ac: add checks for more types, functions and headers makepkg: deal with variable substitution when checking sanity Print callback messages to stderr pacman-key: rework importing distro/repo provided keyrings pacman-key: import everything then revoke on --populate pacman-key: update trust database for relevant operations pacman-key: Improve documentation for --populate pacman-key: rework and document holding keys in keyring Run pacman test-suite with LC_ALL=C Remove outdated comment for _alpm_outerconflicts Dan McGee (148): scripts/pkgdelta: fix `make distcheck` pacman/query.c: avoid variable redeclaration Update MD5 routines with changes from PolarSSL Update base64 PolarSSL code Add sha2 (sha256) routines from PolarSSL Add ALPM sha256sum routines Load and allow access to sha256sum Remove checksum access indirection decode_signature: guess signature data length for efficiency Show 'None' in Signatures -Qip/-Si output if none found Validate the sha256sum if available Parse conflicts/provides/replaces at database load time Add short example signature config to pacman.conf Further fixes to replacement iteration Only check necessary signatures and checksums repo-add: reorganize output messages for clarity Update translations from Transifex Update message catalogs Update authors and contributors doc/PKGBUILD: fix Asciidoc formatting issues Fix compilation using --without-gpgme Fix trailing whitespace in whole codebase Enhance and utilize database status flags Remove use of no-op accessor functions in library Remove usages of alpm_list_next() in backend Replacements refactor: extract check_literal() Replacements refactor: extract check_replacers() Add an _alpm_depcmp_literal() function Fix replacement of provider issue Convert resolvedep() to use _alpm_depcmp_literal() doc/PKGBUILD: update regarding versioned package fields Be more robust when copying package data Rework finding a writable cache directory Remove alpm_option_get_* usage from backend pacsort: ensure list is freed if size is 0 Print debug timestamps to same stream as rest of output dload: prevent need to copy struct in mask_signal() Retrieve default SigLevel value from backend after initial setting Style-match rpmvercmp code with upstream vercmp: ensure 2.0a and 2.0.a do not compare equal Update existing translations from Transifex Add new 'lt' and 'zh_TW' translations from transifex Use off_t rather than double where possible Finish large file download attack prevention Add a few more sizes to humanize_sizes() Slight refresh of the download progress callback Remove short/long label distinction Remove trans is NULL check in QUESTION/EVENT/PROGRESS macros Remove argument from check_pgp_helper strtrim: don't move empty string Fix small memory leak in sig check code Always process validity value returned by gpgme Fix possible mismatched type with several curl arguments Make sync error message smarter on unfound targets Check local database status flag in db_write sanity check block Refactor signature result return format Add more info to debug key display Refactor some args out of pkg_load_internal Better error handling out of package load method pacman-key: refine permission and locking checks pacman-key: have --init add more options to default gpg.conf Allow access to package origin data Database read optimizations _alpm_splitdep(): don't pass bogus length value to strndup Merge branch 'allan/pacman-key' Rework scriptlet tests to not depend on pacman output Streamline alpm_splitdep() comparisons pacman-key: fix broken logic around updating trust database Ensure progress callback updates if XX/YY numerator changes Parse > 2GiB file sizes correctly Scale package integrity progress bar/percentage by package size pacman-key master key generation pacman-key: add --lsign-key operation makepkg: fix overzealous PGP signature file matching Add pactest for overflowing date (year 2038 problem) _alpm_parsedate(): return time_t and not long pactest: remove PACMAN_OUTPUT from ldconfig tests Clean up handling of size fields pacman-key: document --lsign-key doc: consistency when referencing other options pacman-key help and documentation cleanup pacman-key: split keyserver to a separate option pacman-key: add --refresh-keys operation Make delta validation/application more logical Be fully silent on any -Sp operation Move all callbacks up to the handle level Former transaction callback rename refactor Add -p/--print tip for -Q operations on filepath More package operations cleanup dload: use intmax_t when printing off_t Ensure PackageRequired works as expected Remove unnecessary cast query check: use provided filelist count instead of keeping track Use more correct integer types in diskspace checks Print 'loading packages' message on -U operations Move download callback static vars into function download callback: show decimal places in rate if we have room Don't duplicate header strings Fix build without gpgme Ensure entire struct is zeroed in _alpm_parsedate() Add a random pactest Remove dead changelog_feof() code Remove const specifier from changelog_read() void parameter be_local: cope with a desc file without trailing newline Search for non-prefixed paths in skip_remove list Be smarter about running ldconfig during removal transactions Reduce path allocation on the stack in local database Use more efficient way of restoring working directory Flip getcwd()/chdir() for open()/fchdir() in the frontend Access db->pkgcache directly in db_free_pkgcache() _alpm_runscriptlet(): remove clean_tmpdir variable Remove all usages of alpm_list_getdata() from the library utils/cleanupdelta: remove unneeded syncdbpath pacman: use dynamic string allocation where it makes sense Extract an _alpm_pkg_validate_internal() method Split package validation and load loops Allow our PGP helper method to pass back the signature results signing: add a process and retry loop for database signatures Pass package signature data up one more level _alpm_sync_commit: extract two static methods Fix int/size_t type in alpm_list_count() call Remove noisy debug logger Fix possible segfault if siglist was empty validate_deltas: split verify/check errors loops Move default siglevel value from backend to frontend Add alpm_capabilities() method and enumeration Check capabilities in SigLevel option parsing Pull option names out of messages in config parsing config parsing: include file and line number in more errors config parsing: add note if libcurl disabled and no XferCommand Fix signature printing in package info Add new import key question enum value and stub frontend function Add _alpm_process_siglist() logic to failed package validation signing: document most undocumented functions Implement PGP key search and import pacman-key: ensure array iterations are quoted pacman-key: clean up populate output pacman-key: print message in populate if signature is missing pacman-key: don't hide --verify details in populate pacman-key: simplify import in populate Add status and check for key being disabled pacman-key: disable, don't remove, revoked keys pacman-key: implement promptless lsigning pacman-key: remove holdkeys functionality pacman-key: factor out validate_with_gpg() method pacman-key: add an additional plain text 'foo-trusted' file Translation file updates from Transifex Regenerate message catalogs and translations Dave Reisner (67): lib/dload: avoid deleting .part file on too-slow xfer pactest/sync200: check for curl instead of fetch add pactest for replacing a shared provider alpm.h: fix inconsistency in function prototype makepkg: don't hardcode path to strip README: update with 3.5.0 -> 4.0.0 API changes pacman/package: show presence of signature in in -Si lib/sync: reset flag after rejecting a replace dist: preserve symlinks on installation build-sys: always use $(RM) instead of rm -f lib/sync: ignore download errors for as long as possible dload: zero out pm_errno in curl_download_internal lib/dload: avoid renaming download to 0 length destfile lib/dload: refactor deletion on failure sync: fix garbled output in conflict prompt lib/dload: prevent possible NULL dereference makepkg: quote re-evaluation of simple vars dload: add 'unlink_on_fail' to payload struct lib/dload: unlink on response code >=400 bash_completion: add dbonly to sync options dload: remove code duplication dload: rename payload->filename to payload->remote_name dload: delete zero length downloads on curl error dload: remove unused macro pacman/util: use string_length to calculate line length dload: move tempfile and destfile to payload struct dload: rename cd_filename to content_disp_name dload: add open_mode to payload struct dload: move curl option setting to static function dload: move (un)masking of signals to separate functions dload: extract tempfile creation to its own function sync: halt file discovery if repo has no servers pacsort: use boolean, not binary OR operation makepkg: fix breakage in eval'ing quoted strings makepkg: read filenames in a while loop makepkg: fix sanity checking in versioning pacman/callback: reuse strlen calculation avoid blowing out the filecache list when using tmpdir sync: add missing newline in warning message dload: improve debug output dload: abstract dload_interrupted reasons rankmirrors: properly sort resulting times makepkg: unset variables as per !{make,build}flags makepkg: use globs in place of regex makepkg: refactor check_pgpsigs output makepkg: act on function return value, not output makepkg: avoid for loop in deleting manpage hardlinks makepkg: use more awk'ish syntax in sanity checks diskspace: fix memory leak on root mount not found paccache: escape . in trimming of diskspace string makepkg: unset errexit when sourcing /etc/profile dload: provide optional netrc support make compute_download_size consider .part files account for partial delta files in download size makepkg: check for var existance before file existance makepkg: add missing newline on passing gpg sourcecheck makepkg: fix recreation of hardlinks to .gz manpages buildsys: remove existing symlinks before installing include ldconfig.stub in EXTRA_DIST pacman: add short opt '-p' for --print to -{S,R}h paccache: remove unnecessary if check dload: remove user:pass@ definition from hostname configure: Fix quoting in SEDINPLACE on Darwin pactree: include missing ctype.h header dload: avoid using memrchr src/util: provide strndup definitions where needed pacman/upgrade: print 'loading packages...' only once Diogo Sousa (3): Improved alpm_list_mmerge() performance (fixed coding style) lib/libalpm/signing.c: Fix memory leak in decode_signature() in case of error. Removed multiple definitions of pkgfrom_t Helder Martins (1): lib/libalpm/handle.c: Removed redundant if condition Jakob Gruber (1): Fix formatting in pacman-key manpage Lukas Fleischer (7): Skip rename() on NULL destfile in curl_download_internal() Avoid stat() on NULL path in curl_download_internal() lib/libalpm/dload.c: Add ASSERT() to alpm_fetch_pkgurl() lib/libalpm/dload.c: Use STRDUP() instead of strdup() pacman.c: Remove redundant strdup() in parsearg_global() be_sync.c: Fix memory leak in alpm_db_update() lib/libalpm/util.c: Fix two memory leaks Pang Yan Han (4): pactest: -U --recursive pacsort: correct pointer type in list_new pacsort: correct list freeing Improve advice for sync ops when db.lck is present Rémy Oudompheng (1): libalpm: fix a remaining old syntax of RET_ERR() macro Sebastien Luttringer (2): Update PKGBUILD manpage about startdir deprecation Add makepkg -S which is an alias to makepkg --source ----------------------------------------------------------------------- hooks/post-receive -- The official pacman repository

1 0