I've made an attempt to add ETA functionality to `qlop --current` The algorithm is the same as genlop uses with one minor difference. If the user is allowed to read /proc/pid/environ, it will fetch the category from it to work around ambiguous packagenames. The average time calculation is a stripped version of `qlop --time`. All timestamps are read from emerge.log and the last match is the starttime for the current emerge. I found this easier to work with then the process uptime. If you like it you're free to include it, but feedback is always welcome. Thanks for these excellent tools :) Harm Geerts
Created attachment 106286 [details, diff] portage-utils-0.1.22-qlop-eta.patch
I'm getting failures on testing your patch local. Would you mind re doing your patch to cvs vs the last release. fbsd added some code to this same function. Infact some of the code you have reword could/should probably be split out into a new function so less code would have to be duped.
This is gonna get in a conflict with the change I've committed to support qlop -c on FreeBSD. I think the cleanest way to handle this is to factor out the process data gathering (platform dependent) from the displaying and the ETA support, so that the platform specific function would just get the start date and the time elapsed, while the platform-independent code can take care of the rest. This way, we wouldn't need to copy the code for every new platform supported (almost every platform has its different interface to get processes' data), but would just have to find the two parameters. Actually there's just one information needed, the start_date, as you can easily find the elapsed seconds since that once you finished the processing (that might take a few seconds, I suppose, for big enough logfiles).
Created attachment 106430 [details, diff] portage-utils-cvs-qlop-eta.patch no fair, I submitted my patch before you committed yours ;) Couldn't test this on *BSD, but I think I got it right. The start_date is the last match from emerge.log, it reflects when emerge started working on a package. So all we need to get the correct result is the packagename. I think this is better then pulling the start_date from the process info, which only tells us the elapsed time of the current stage. (unpack, compile, etc.)
Created attachment 106442 [details, diff] portage-utils-cvs-qlop-eta.patch *cough* shouldn't have used cvs diff
Flameeyes: Please cvs up and test this on fBSD. tia
Pushed 0.1.23 but excluded this patch. needinfo from bsd team.
Something is not right with your code. pitr ~ # ./q lop -c Segmentation fault pitr ~ # q lop -c * desktop-file-utils-0.11 started: Fri Jan 19 19:39:54 2007 elapsed: 12 seconds pitr ~ # ./q lop -c Segmentation fault The q in ./ is with your additions. (gdb) bt full #0 0x08074321 in strstr () No symbol table info available. #1 0x08054ac5 in show_current_emerge (logfile=0x80d9ffc "/var/log/emerge.log") at qlop.c:429 proc = (DIR *) 0x80d9f24 de = <value optimized out> buf = {"[nautilus-2.16.3] sandbox\000/usr/lib/portage/bin/ebuild.sh compile", '\0' <repeats 8127 times>, "PORTAGE_DEPCACHEDIR=/var/cache/edb/dep\000LESS=-R -M --shift 5\000OPENGL_PROFILE=xorg-x11\000DISTDIR=/var/tmp/portage/nautilus-2.16.3/distdir\000INFOPATH=/usr/share/info:/usr/s_64-pc-linux-gnu/3."...} path = "/proc/13028/environ\000\000\000\000\000\000\000\200\206\020\bØú\204ÿsÆ\b\b\002\000\000\000ø\r\205ÿÑ\237\r\b" p = 0xff84baaa "] sandbox" atom = <value optimized out> #2 0x080558e8 in qlop_main (argc=2, argv=0xff850df8) at qlop.c:539 i = <value optimized out> average = 1 do_time = 0 '\0' do_list = 0 '\0' do_unlist = 0 '\0' do_sync = 0 '\0' do_current = 1 '\001' do_human_readable = 0 '\0' opt_logfile = 0x0 logfile = 0x80d9ffc "/var/log/emerge.log" __FUNCTION__ = "qlop_main" #3 0x0804ea25 in q_main (argc=3, argv=0xff850df4) at q.c:132 buf = '\0' <repeats 3772 times>, "$\234\r\b\016\000\000\000\001\000\000\000 \230\017\b\000\002\000\0003\234\r\b\f\000\000\000\002\000\000\000 \232\017\b\000\004\000\000@\20\000\001\000\000\0003\024\020\b\024\000\000\000E\234\r\b\016\000\000\000\001\000\000\000àÀ\017\b\000\020\000\000¿\231\r\b\a\000\000\000\000\000\000\000k\v\205ÿ\001\000\000\000T\2000\002\000\000\000 \226\017\b\000\002\000\000]\234\r\b\a\000\000\000\001\000\000\000@ \017\b\000\020\000\000e\234\r\b\017\000\000\000\001\000\000\000 \224\017\b\000\002\000\000u\00\000\001\000\000\000 \222\017\b\000\002\000\000\204\234\r\b\006\000\000\000\001\000"... i = <value optimized out> p = <value optimized out> func = (APPLET) 0x80555f0 <qlop_main> __FUNCTION__ = "q_main" #4 0x0805e11e in main (argc=3, argv=0xff850df4) at main.c:1035 No locals. (gdb) bt #0 0x08074321 in strstr () #1 0x08054ac5 in show_current_emerge (logfile=0x80d9ffc "/var/log/emerge.log") at qlop.c:429 #2 0x080558e8 in qlop_main (argc=2, argv=0xff850df8) at qlop.c:539 #3 0x0804ea25 in q_main (argc=3, argv=0xff850df4) at q.c:132 #4 0x0805e11e in main (argc=3, argv=0xff850df4) at main.c:1035
I'm a little confused by your backtrace. The bt is from a segfault on line 429 It says p = 0xff84baaa "] sandbox" which is from buf[0] But on line 428 I assign p to buf[1] Why does p still point to buf[0] ? 424: if (buf[0][0] == '[' && (p = strchr(buf[0], ']')) != NULL && strstr(buf[0], "sandbox") != NULL) { 425: /* try to fetch category from process env */ 426: snprintf(path, sizeof(path), "/proc/%i/environ", pid); 427: if (eat_file(path, buf[1], sizeof(buf[1]))) { 428: p = buf[1]; 429: while (strstr(p, "PORTAGE_BUILDDIR=") == NULL) 430: p = strchr(p, '\0')+1;
I must of had other changes pending or so as I'm unable to reproduce the error now /var/cvsroot/gentoo-projects/portage-utils/qlop.c,v <-- qlop.c new revision: 1.38; previous revision: 1.37
The bug is valid, but I'm not sure how to solve it in a decent manner. It triggers when /proc/PID/environ is 4096 bytes and doesn't hold the variable PORTAGE_BUILDDIR 4096 bytes seems to be the maximum and the rest is trimmed off. I made a hack that searches multiple variables until it finds a match but even then it is possible for this to trigger so I didn't want to include that either. I didn't expect this to be included which is why I haven't responded until now. It's probably best to take out the part that searches for the category in /proc/PID/env
Ok then I've reverted the patch. If/when you have a new one that's bug free feel free to attach it here for inclusion in a future release. /var/cvsroot/gentoo-projects/portage-utils/qlop.c,v <-- qlop.c new revision: 1.39; previous revision: 1.38
Reopen when ready.
Any news on this? This is the main reason for to still have genlop installed... Reopening this bug would be nice.
(In reply to comment #14) > Any news on this? This is the main reason for to still have genlop > installed... Reopening this bug would be nice. ETA calculation is still missing in qlop. This issue is definitely not RESOLVED/FIXED.
I find the proposal of Bug 498338 rather convincing. I.e. we should not include the ETA per default (as it requires scanning the complete emerge.log file and thus is expensive if invoked on a regular basis. E.g. i use qlop in conky), but allow the combined parameter -ct to show the expected merge time in addition to the current package build time.
*** Bug 498338 has been marked as a duplicate of this bug. ***
*** Bug 569104 has been marked as a duplicate of this bug. ***
I'd love seeing this implemented :) qlop is generally way faster than genlop
I'm not that optimistic it would be easy to get here, but technically speaking from the log we'd be able to harvest unfinished merges - that as a start
The bug has been referenced in the following commit(s): https://gitweb.gentoo.org/proj/portage-utils.git/commit/?id=c74a5abf3fd8db03adb531f95ecff5316d997ab3 commit c74a5abf3fd8db03adb531f95ecff5316d997ab3 Author: Fabian Groffen <grobian@gentoo.org> AuthorDate: 2019-02-27 20:41:45 +0000 Commit: Fabian Groffen <grobian@gentoo.org> CommitDate: 2019-02-27 20:50:21 +0000 qlop: rewrite from scratch This new implementation achieves a few things: - unified code path for all modes of operation (thus consistent results) - more flexibility to implement new features - simplification of the codebase Short version of what this commit changes: - existing flags -t -g have been replaced with -a and -t - -c has been been replaced with -r and no longer uses proc processing code (thus works everywhere) - addition of an ETA for running emerges (subject to improvements) - allow reading a file of atoms (e.g. /var/lib/portage/world) - summary mode -c to compute grand total, e.g. to compute world compile time Bug: https://bugs.gentoo.org/161244 Bug: https://bugs.gentoo.org/603024 Bug: https://bugs.gentoo.org/636334 Bug: https://bugs.gentoo.org/658824 Signed-off-by: Fabian Groffen <grobian@gentoo.org> man/include/qlop.desc | 25 +- man/include/qlop.optdesc.yaml | 23 +- man/qlop.1 | 66 +- qlop.c | 1479 +++++++++++++++++++++-------------------- tests/qlop/dotest | 20 +- tests/qlop/list01.good | 4 +- tests/qlop/list02.good | 6 +- tests/qlop/list03.good | 8 +- tests/qlop/list04.good | 2 +- tests/qlop/list05.good | 2 +- tests/qlop/list06.good | 4 +- tests/qlop/list07.good | 4 +- tests/qlop/list08.good | 4 +- tests/qlop/list09.good | 6 +- 14 files changed, 892 insertions(+), 761 deletions(-)
this is available onwards from 0.80_pre*
