And this is without your PR with dropping caches:
```
⋊> ~/P/u/rpm on master ⨯ time ./rpm -Va >/dev/null
15:29:57
Command exited with non-zero status 1
53.21user 12.86system
With your PR (with/without dropping caches):
```
⋊> ~/P/u/rpm on master ⨯ time ./rpm -Va >/dev/null
15:26:03
Command exited with non-zero status 1
297.33user 12.27system
@pmatilai pushed 1 commit.
b5da21861ad70b9df5c2cd5560a921251a5778d7 fixup! Run file verification in
parallel threads
--
You are receiving this because you are subscribed to this thread.
View it on GitHub:
Hmm, couple of tests are failing:
1) because errno isn't stored in the verify results, but that only affects the
produced messages
2) ghost etc skipping isn't handled in the second loop
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it
If you could provide some script for testing this, I can do some benchmarks.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
While we're playing with parallelising things...
On my rusty old laptop, this tends to just slow things down unless
things are hot in cache, in which case it roughly halves the running time.
Might be worth more on SSD based setup, I'd be interested to hear results on
such systems.
For now, this
Dunno what there is to script. Try 'time rpm -Va' with and without, add "echo 1
> /proc/sys/vm/drop_caches" in between to see what happens with cold caches?
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
Thing is, for whatever reason packagers are keen to do whatever crazy things
they can get away with in specs, often for no apparent reason. We don't change
rpm just to appease to that stuff.
Patch to document the intended and implemented behavior is welcome, and the
missing \n in the message
No further action -> closing
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/issues/671#issuecomment-491262095___
Rpm-maint mailing
Closed #671.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/issues/671#event-2332894385___
Rpm-maint mailing list
In the comment your patch removed:
/* Print failure iff %{load:...} or %{!?load:...} */
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
Please where is this usage of %{?load:...} described (not counting macro.c)?
I see %{?load:...} in spec files ruby.spec or vagrant.spec where it is used for
another purpose: to be able to build SRPM on older Fedora, where the older RPM
has not %load defined. And it looks that the packager
NAK.
As the comment indicates, this behavior is intentional to optionally allow
silent failure, eg to try to load a macro file that is not there in all cases.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
Closed #702.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/702#event-2332827595___
Rpm-maint mailing list
@pmatilai pushed 1 commit.
ec033f4329e404e8d37e8de804e3fe1006a6536f Eliminate the hysterical filesize
retval from rpmDoDigest() now that we can
--
You are receiving this because you are subscribed to this thread.
View it on GitHub:
Doh, and another force-push to eliminate redundant Ftell()
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
Forgot to eliminate libelf dependencies in the first round, hence the forced
push.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
Back in the turn of the century somebody thought it was a neat idea
to completely compromise system security to improve program start-up
start-up times a wee bit. Since then, people have thankfully started
coming to their senses and removed prelink from distros entirely.
Lets stop papering over
The patch aside (which I believe is necessary at any rate), discussion on what
to do about this would be welcome. RHEL carries a patch to enable a "strict
mode" for scriptlet failures (ie bring back the duplicates on
scriptlet-failures), and Suse seems to have something similar.
--
You are
When called in spec context, the package structs are properly freed
but in rpmdeps context, commit 49f2bb7d8fd91f2d8b22bf7128fd3defe4ed5434
only added a "dirty kludgery" to make it not blow up. This causes
the rpmds structures created in rpmfcApplyInternal() to leak memory.
Make freePackage()
This alleged solution to the problem of not all scriptlet errors being
reflected in our exit code is not a solution at all, it just pushes
the problem to somebody elses lap and introduces inconsistency and
an untracked incompatibility as well.
This reverts commit
Always check for background first. This appears to have been accidentally added
in commit 9343ecd94cd873e6dc1c06428975163cbb9cf9af, which should be mentioned
in the commit message. Other than that it's obviously fine.
--
You are receiving this because you are subscribed to this thread.
Reply
In the commit to pass rpmts object ro rpmSpecBuild() there's a stray empty line
added to build/build.c, but it seems to get lost somewhere in the birds-eye
view that GH considers review mode.
The birds-eye view of things is useful for, well, a birds-eye view, but
bulls*** for reviewing patch
pmatilai commented on this pull request.
> if (rstreqn(N, "rpmlib(", sizeof("rpmlib(")-1)) {
- if (tagN != RPMTAG_REQUIRENAME) return 1;
+ if (tagN != RPMTAG_REQUIRENAME &&
+ (tagN == RPMTAG_PROVIDENAME && !(Flags & RPMSENSE_RPMLIB)))
+ return 1;
Fix the
pmatilai commented on this pull request.
> rpmlog(RPMLOG_NOTICE, _("\n\nRPM build errors:\n"));
rpmlogPrint(NULL);
}
rpmugFree();
-
+if (missing_buildreqs && !rc) {
+ rc = RPMRC_MISSINGBUILDREQUIRES;
+}
+if (rc == RPMRC_FAIL)
+ rc = 1;
Please use
pmatilai commented on this pull request.
> @@ -273,20 +287,6 @@ static struct poptOption optionsTable[] = {
POPT_TABLEEND
};
-static int checkSpec(rpmts ts, rpmSpec spec)
-{
-int rc;
-rpmps ps = rpmSpecCheckDeps(ts, spec);
-
-if (ps) {
- rpmlog(RPMLOG_ERR, _("Failed
pmatilai commented on this pull request.
> @@ -273,20 +287,6 @@ static struct poptOption optionsTable[] = {
POPT_TABLEEND
};
-static int checkSpec(rpmts ts, rpmSpec spec)
-{
-int rc;
-rpmps ps = rpmSpecCheckDeps(ts, spec);
-
-if (ps) {
- rpmlog(RPMLOG_ERR, _("Failed
pmatilai commented on this pull request.
> @@ -894,6 +895,9 @@ static rpmSpec parseSpec(const char *specFile,
> rpmSpecFlags flags,
case PART_PREP:
parsePart = parsePrep(spec);
break;
+ case PART_BUILDREQUIRES:
+ parsePart =
pmatilai commented on this pull request.
> goto exit;
+ if (what & RPMBUILD_BUILDREQUIRES)
+rc = doBuildRequires(spec, test);
+ if ((what & RPMBUILD_CHECKBUILDREQUIRES) &&
+ (rc == RPMRC_MISSINGBUILDREQUIRES))
+ rc =
pmatilai commented on this pull request.
> +/* add results to requires of the srpm */
+argvSplit(, getStringBuf(sb_stdout), "\n\r");
+outc = argvCount(output);
+
+if (!outc) {
+ rc = RPMRC_OK;
+ goto exit;
+}
+
+for (int i = 0; i < outc; i++) {
+
pmatilai commented on this pull request.
> @@ -213,24 +278,58 @@ static rpmRC buildSpec(BTA_t buildArgs, rpmSpec spec,
> int what)
} else {
int didBuild = (what & (RPMBUILD_PREP|RPMBUILD_BUILD|RPMBUILD_INSTALL));
+ if (!spec->buildrequires && (what &
pmatilai commented on this pull request.
> @@ -16,6 +16,8 @@
#include "debug.h"
+#define RPMRC_MISSINGBUILDREQUIRES 11
Hmm, we're actually returning this in the API so I think it should actually go
into rpmbuild.h
--
You are receiving this because you are subscribed to this thread.
pmatilai commented on this pull request.
> +for (int i = 0; i < outc; i++) {
+ parseRCPOT(spec, spec->sourcePackage, output[i], RPMTAG_REQUIRENAME, 0,
0, addReqProvPkg, NULL);
+}
+
+rpmdsPutToHeader(*packageDependencies(spec->sourcePackage,
RPMTAG_REQUIRENAME),
pmatilai commented on this pull request.
> @@ -172,9 +179,67 @@ rpmRC doScript(rpmSpec spec, rpmBuildFlags what, const
> char *name,
return rc;
}
-static rpmRC buildSpec(BTA_t buildArgs, rpmSpec spec, int what)
+static rpmRC doBuildRequires(rpmSpec spec, int test)
+{
+StringBuf
pmatilai commented on this pull request.
> +
+if (!spec->buildrequires) {
+ rc = RPMRC_OK;
+ goto exit;
+}
+
+if ((rc = doScript(spec, RPMBUILD_BUILDREQUIRES, "%generate_buildrequires",
+ getStringBuf(spec->buildrequires), _stdout, test)))
+
pmatilai commented on this pull request.
> @@ -283,7 +284,7 @@ static rpmRC buildSpec(BTA_t buildArgs, rpmSpec spec, int
> what)
return rc;
}
-rpmRC rpmSpecBuild(rpmSpec spec, BTA_t buildArgs)
+int rpmSpecBuild(rpmts ts, rpmSpec spec, BTA_t buildArgs)
Just noting that this is an API
pmatilai commented on this pull request.
> @@ -172,6 +172,7 @@ rpmRC doScript(rpmSpec spec, rpmBuildFlags what, const
> char *name,
return rc;
}
+
Stray newline added.
Like said in an earlier comment, the addition of rpmtsFromPyObject() could be
merged with this commit because it
pmatilai commented on this pull request.
> @@ -47,7 +47,7 @@ static rpmRC doRmSource(rpmSpec spec)
* @todo Single use by %%doc in files.c prevents static.
*/
rpmRC doScript(rpmSpec spec, rpmBuildFlags what, const char *name,
- const char *sb, int test)
+ const
Argh, I didn't mean to approve the whole thing, thought I could do it for the
couple of earlier commits separately.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
pmatilai approved this pull request.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/rpm-software-management/rpm/pull/593#pullrequestreview-235945251___
40 matches
Mail list logo
