The problem here is that the cat is out of the bag and people are already using
whatever in their specs. This is a bunch of random ponderings on the subject,
rather than anything definitive on way or the other.
If we limited it to UTC only, then the "UTC" part becomes waste of perfectly
good
Dropping 4.15 milestone, this is just any old bug that's a regression and not
release critical, and sorting out apparently needs more time than we have.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
This is turning into a whole lot of pain for bleeping little gain. Rpm is not
in the business for its timezone knowledge, so we'll certainly not be
incorporating our own timezone tables to accommodate for random timezone
abbreviation preferences folks happen to have.
--
You are receiving this
@pavlinamv Australian official eastern time designation is `AEST`. Australians
do know this, and will use it when appropriate.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
@pavlinamv The `date` command can do exactly the same thing as this.
```
$ date --date='Wed 31 Jul 2019 09:02:40 AM EDT' +'%s'
1564578160
$ date --date='@1564578160'
Wed 31 Jul 2019 09:02:40 AM EDT
$ date --date='@1564578160' --utc
Wed 31 Jul 2019 01:02:40 PM UTC
```
Cf.
>@pavlinamv So then, how does date in GNU Coreutils do it?
rpm task:
_input_: time in an arbitrary time zone, the time zone description
_result_: the corresponding time in UTC.
Command ```date``` _prints_ or _sets_ the system date and time. Neither
_printing_, nor _setting_ is so complex (or
@pavlinamv So then, how does `date` in GNU Coreutils do it?
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
>There is no reason to break support for this,
@Conan-Kudo, @ignatenkobrain:
The current situation is:
rpm accepts arbitrary text as a time zone description in extended date format.
If the time zone description is not "correct" rpm (silently) change it to UTC
instead of the given time zone
FYI, I do use CET vs CEST in all spec files. Don't break that :)
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
@pavlinamv It is not true that the timezone short codes do not encode whether
daylight saving time/summer time is in effect.
For example, where I live, daylight saving time means Eastern Time is
represented with `EDT`, whereas standard time is represented as `EST`.
Europe has `CET` vs `CEST`,
pmatilai commented on this pull request.
> @@ -44,7 +44,8 @@ static int dateToTimet(const char * datestr, time_t * secs,
> int * date_words)
struct tm time, ntime;
const char * const * idx;
char *p, *pe, *q, *date, *tz;
-char tz_name[10]; /* name of timezone
pmatilai commented on this pull request.
> @@ -178,13 +179,17 @@ static int dateToTimet(const char * datestr, time_t *
> secs, int * date_words)
if (*secs == -1) goto exit;
+if (tzname[1][0] == 0)
+ rpmlog(RPMLOG_WARNING, _("time zone in %%changelog not in tz database:
pmatilai commented on this pull request.
> @@ -44,7 +44,8 @@ static int dateToTimet(const char * datestr, time_t * secs,
> int * date_words)
struct tm time, ntime;
const char * const * idx;
char *p, *pe, *q, *date, *tz;
-char tz_name[10]; /* name of timezone
pavlinamv commented on this pull request.
> @@ -178,13 +179,17 @@ static int dateToTimet(const char * datestr, time_t *
> secs, int * date_words)
if (*secs == -1) goto exit;
+if (tzname[1][0] == 0)
+ rpmlog(RPMLOG_WARNING, _("time zone in %%changelog not in tz database:
pavlinamv commented on this pull request.
> @@ -44,7 +44,8 @@ static int dateToTimet(const char * datestr, time_t * secs,
> int * date_words)
struct tm time, ntime;
const char * const * idx;
char *p, *pe, *q, *date, *tz;
-char tz_name[10]; /* name of timezone
pmatilai commented on this pull request.
> @@ -44,7 +44,8 @@ static int dateToTimet(const char * datestr, time_t * secs,
> int * date_words)
struct tm time, ntime;
const char * const * idx;
char *p, *pe, *q, *date, *tz;
-char tz_name[10]; /* name of timezone
pmatilai commented on this pull request.
> @@ -178,13 +179,17 @@ static int dateToTimet(const char * datestr, time_t *
> secs, int * date_words)
if (*secs == -1) goto exit;
+if (tzname[1][0] == 0)
+ rpmlog(RPMLOG_WARNING, _("time zone in %%changelog not in tz database:
pavlinamv commented on this pull request.
> @@ -178,13 +179,17 @@ static int dateToTimet(const char * datestr, time_t *
> secs, int * date_words)
if (*secs == -1) goto exit;
+if (tzname[1][0] == 0)
+ rpmlog(RPMLOG_WARNING, _("bogus TZ database name in %%changelog:
%s\n"),
pmatilai commented on this pull request.
> @@ -178,13 +179,17 @@ static int dateToTimet(const char * datestr, time_t *
> secs, int * date_words)
if (*secs == -1) goto exit;
+if (tzname[1][0] == 0)
+ rpmlog(RPMLOG_WARNING, _("bogus TZ database name in %%changelog:
%s\n"),
Warning instead of info is emitted (in the latest version).
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
Corrected in the latest version.
--
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/739#issuecomment-501619925___
Rpm-maint mailing
Looks good. Only thing left is a missing "g" in the commit message of 3rd patch.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
Panu's comments incorporated.
Moreover I think about the problem again. The current implementation does not
need further changes, but it needs to correct comments and add an info if the
description of time zone is not correct. o the first commit in the PR is
omitted and commit "Emit info if
pmatilai commented on this pull request.
> @@ -130,8 +130,7 @@ static int dateToTimet(const char * datestr, time_t *
> secs, int * date_words)
SKIPNONSPACE(pe);
if (*pe != '\0')
*pe++ = '\0';
- if (((int)(pe-p) + 1) > 9 )
- goto exit;
+ tz_name =
pmatilai commented on this pull request.
> @@ -130,8 +130,7 @@ static int dateToTimet(const char * datestr, time_t *
> secs, int * date_words)
SKIPNONSPACE(pe);
if (*pe != '\0')
*pe++ = '\0';
- if (((int)(pe-p) + 1) > 9 )
- goto exit;
+ tz_name =
pmatilai commented on this pull request.
> @@ -185,6 +184,8 @@ static int dateToTimet(const char * datestr, time_t *
> secs, int * date_words)
rc = 0;
exit:
+if (tz_name)
Don't bother testing for NULL, free() does it anyway so it ends up being just
useless clutter.
--
You are
pavlinamv commented on this pull request.
> @@ -30,10 +30,12 @@ static int sameDate(const struct tm *ot, const struct tm
> *nt)
ot->tm_wday == nt->tm_wday);
}
+#define TZ_MAX_LENGTH 80
Thank you. In the new version dynamic allocation is used.
--
You are receiving this because
pmatilai commented on this pull request.
> @@ -30,10 +30,12 @@ static int sameDate(const struct tm *ot, const struct tm
> *nt)
ot->tm_wday == nt->tm_wday);
}
+#define TZ_MAX_LENGTH 80
Just use dynamic allocation to avoid arbitrary limits that need explanations :)
--
You are
ignatenkobrain commented on this pull request.
> @@ -30,10 +30,12 @@ static int sameDate(const struct tm *ot, const struct tm
> *nt)
ot->tm_wday == nt->tm_wday);
}
+#define TZ_MAX_LENGTH 80
It would be great if this number was described. I mean why not 67 or 93 :)
--
You are
Tests were added. Moreover they already help to show, that exit status 0 is
returned when changelog order fails. Thus I add an additional commit that
corrects this behaviour (regression).
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view
@pavlinamv pushed 2 commits.
1b44c113b06574e24ae5cd1e92aa2ae132975081 Return non-zero exit status if
changelog order fails
e4dc76e450dd0277fe5f15d4b74df3935dc2ae58 Test effects of time zone in chanelog
timestamp
--
You are receiving this because you are subscribed to this thread.
View it
it would be great to get it as a test so we don't break it at some point
accidentally.
--
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
When building RPMs that have %changelog sections with changelog entries with
full timestamps, RPM did not take the time zone into account. Now the timezone
description is taken into account using function tzset(). It handles correctly
timezone descriptions like:
"Europe/London",
"GMT-5", or
33 matches
Mail list logo
