This patch represents the final resting place for miscellaneous
changes not otherwise encountered or for which the resolution
was incomplete or incorrect.
The gettext documentation leads one to believe that
a printf type call is necessary for gettext string
extraction. That turns out to be misleading and
all one really needs is the runtime gettext string
address resolution.
Thus, we can avoid our original snprintf/strdup
overhead and establish an address for the original
or translated string just by issuing the _() macro.
We create these nls string tables so that:
1) top avoids the overhead of repeated
runtime function calls
2) we can control the order of top's
strings in the .pot file
3) translator comments don't obscure
and clutter the main program
Until this patch, top had used some strings with
special escape sequences to produce colors, normal
text, bold text, etc. They took the following form,
explained by an excerpt from program comments:
...
Our special formatting consists of:
"some text <_delimiter_> some more text <_delimiter_>...\n"
Where <_delimiter_> is a single byte in the range of:
\001 through \010 (in decimalizee, 1 - 8)
and is used to select an 'attribute' from a capabilities table
which is then applied to the *preceding* substring.
...
Unfortunately, these nonprinting values revealed
insurmountable inconsistencies in both the front-end
and back-end translation tools.
The xgettext (extraction) program would take those
special escapes, convert them and then output raw
binary values. Thus the .pot file would contain
lots of unprintable stuff making it unreadable.
If the following was added to po/Makevars, most of
those special escapes would be preserved in their
escape notation:
XGETTEXT_OPTIONS = ... --escape
But two escapes were converted from octal notation
and there was no way to prevent it:
\007 --> \a
\010 --> \b
After a pass through the msginit program, most of
the escapes were reconverted to raw binary values
making translation impossible. There was no
"--escape" option for the back-end programs like
there was for xgettext.
But the real killer was the escape \004, also used
in some of top's special strings. This value would
be silently accepted by xgettext, only to produce
the following fatal error in back-end programs like
msginit, msgfmt and msgen:
.pot:2647: context separator <EOT> within string
To quote from one of the references below:
"Would you create a suite of tools that silently
allow what is destined to become a fatal error
to pass unnoticed?"
So the bottom line was: top's special strings, in
use for the past nine years, had to be redesigned.
References:
http://www.freelists.org/post/procps/procpsng-nls-support,11http://www.freelists.org/post/procps/procpsng-nls-support,14
summary of changes:
. adopted relative paths to 'include' and 'proc'
dirs so that stand alone compiles are made
easier and no one need guess their locations
. corrected several names for enums and macro
usage reflecting fmt vs. txt
. expanded all octal escape sequences to a full
3 digits since one already required 3 digits
. finalized translator hints (for now)
programming note:
as an aside, by not including an argument for
the gettext --add-comments, any preceeding c
style comment will be propagated to the .pot
file, if the gettext macro isn't empty.
/* Need Not Say 'TRANSLATORS' ...
snprintf(buf, sizeof(buf), "%s", _( // unseen
/* Translator Hint: ...
snprintf(buf, sizeof(buf), "%s", _("" // seen
With the privatization of the library xalloc functions,
this patch restores top's previous internal memory logic.
This puts us back on a 3.3.1 (and earlier) footing wherein
memory allocation errors carry these implications:
. if top detected, the terminal will be restored
prior to an abnormal exit and message
. if library detected, the terminal will become
corrupted following the message
To my knowledge, neither type of memory error has ever occurred
during the nine years since top was rewritten. So the issue of
restoring the termios structure is probably moot.
By not limiting --add-comments= via an argument, xgettext is far
too aggressive (one might say stupid) when extracting comments.
It doesn't even limit extraction to a single preceeding comment.
Here is an example showing a program source excerpt and the
resulting .pot excerpt:
--- program source ----------------------------------------
close (0); dup (in_fd[0]); /* set the stdin to the in pipe */
close (1); dup (out_fd[1]); /* set the stdout to the out pipe */
close (2); dup (out_fd[1]); /* set the stderr to the out pipe */
execvp (args[0], args); /* exec gdb */
perror (_("exec failed"));
--- resulting .pot ----------------------------------------
#. set the stdin to the in pipe
#. set the stdout to the out pipe
#. set the stderr to the out pipe
#. exec gdb
#: ps/stacktrace.c:28 ps/stacktrace.c:63
msgid "exec failed"
msgstr ""
. 1 program uses curses (top)
. 2 programs use ncurses (slabtop, watch)
. all 3 include non-wide <curses.h> or <ncurses.h>
. those 3 do not currently need wide support
. but anticipating nls, we link against libncursesw
This patch ensures an environment consistent with current
and future ncurses needs.
This commit will enable pass comments to translators near gettext
directive to .pot file. For example;
/* TRANSLATORS: this is visible in .pot file */
printf(_("to be translated"));
Signed-off-by: Jim Warner <james.warner@comcast.net>
This commit documents the top search provisons as a
new section entitled:
5d. SEARCHING in a window
It also adds minor unrelated content and tweaks some
miscellaneous formatting and spacing inconsistencies.
Commit a5881b5a4e, addressing
command lines with a trailing space, produced an undesirable
side effect in the fill_cgroup_cvt() function.
This patch restores correct cgroup behavior while still
producing command lines with no trailing space.
There were numerous ps memory allocation inconsistencies.
Some were checked for failure and others were not.
The program was modified to utilize the library memory
rouines which are consistent in dealing with errors.
(a few changes simply removed trailing whitespace)
The smatch error --------------
top.c +1414 calibrate_fields(78) error: buffer overflow 'Fieldstab' 39 <= 39
The code -----------------------
if (P_MAXPFLGS < f) { w->endpflg = i; continue; }
The background -----------------
The enum P_MAXPFLGS is strictly a fencepost and can *never* appear in
the arrays pflgsall or procflgs. Thus it (39th element) cannot be used
in referencing Fieldstab.
However, two enums of higher value (X_XON=40 and X_XOF=41) *can* appear
in those arrays. But the test against the fencepost ensures that those
two enums are *never* used in referencing Fieldstab.
When the analyzer sees the conditional using '<' and not '<='
it reports a false positive.
The source was changed to accommodate the tool's deficiency
With forest view mode under procps-ng-3.3.0, top would catch
a SEGV and exit. The problem was thought to be related to
some linux version 3 anamoly.
In procps-ng-3.3.1, top was protected from abnormal exit and
would display the offending processes with the special orphan
notation '?', instead of the usual artwork.
As it turns out, the problem wasn't with linux and could
surface under a 2.6.xx kernel, The cause was an occasional
abnormal ordering of tid/tgid/ppid which went unresolved
due to a progressive proc table scan in the recursive
function forest_add().
By default the proc table scan now restarts at the beginning.
However, a #define TREE_ONEPASS is provided should one wish
to restore the 3.3.1 behavior with its reduced overhead.
An anomaly would surface whenever the sort column was
made non-display under fields management.
The scroll message would report incorrect x coordinates
and the 'x' toggle would produce an unintentional
horizontal scroll.
This patch forces column highlighting off whenever field
changes are made (as is done for horizontal scrolling).
It also inactivates the 'x' toggle whenever the sort
column is not visible.
This bug has yet to be reported.
In the procs_refresh() function:
. the proctable scan was simplified
. memory allocation is now much more efficient
. window ppts will usually not now need xrealloc
Now that the library allows for customizing the response to
memory allocation errors, it is no longer necessary for top
to provide its own.
This library extension was necessary since top modifies the
termios structure and there would otherwise be no way to
restore the users tty.
When PROC_FILLARG was used (invoking file2strvec)
command lines contained no trailing space.
When PROC_EDITCMDLCVT was used (invoking read_unvectored)
command lines contained a trailing space.
Now both routes to a cmdline act the same -- no trailing space.
The CodingStyle document suggests programs should allow for
cmdlines of at least 128k. Only the ps program can display
such a cmdline, and only with multiple -w switches.
The library function file2strvec can essentially return
a cmdline of unlimited length. However, the library function
fill_cmdline_cvt used an arbitrary upper limit of 2048 for
buffers automatically allocated on the stack.
This patch raises the fill_cmdline_cvt upper limit to 128k via
dymaically acquired utility buffers ensured by the openproc
function.
It also makes indentation consistent in the openproc function.
There were numerous library memory allocation inconsistencies.
Some were checked for failure and others were not.
All library source modules were modified to utilize the alloc.h
memory rouines which are consistent in dealing with errors.
A callback provision in the form of xalloc_err_handler
(of type message_fn) was added to the alloc module.
This change allowed a program like top, who alters the
termios structure, to override the default fprint(stderr...)
behavior in the event of an error.
The new function xstrdup was also added for symmetry.
. 3 pgms require non-wide <curses.h> or <ncurses.h>
This patch represents the tests for a minimal environment
consistent with current ncurses needs.
It should allow a successful configure and build
. 1 program uses curses (top)
. 2 programs use ncurses (slabtop, watch)
. all 3 include non-wide <curses.h> or <ncurses.h>
. those 3 do not currently need wide support
. but anticipating nls, we link against libncursesw
This patch ensures an environment consistent with current
and future ncurses needs.
Program changes
. corrected eolcap handling if EQUCOLHDRYES defined
. added missing terminator for OOMEM_ENABLE comment
. tightened logic in display_fields
. eliminated some warnings via type or cast
. updated miscellaneous source comments
Man Document changes
. corrected/expanded CGROUPS narrative
. standardized/normalized PPID narrative
Signed-off-by: Jan Görig <jgorig@redhat.com>
This source patch addresses the following:
Enhancements
. added -w command line switch, with optional argument
( offers complete control over width and height, in )
( combination with env variables COLUMNS= and LINES= )
. updated man document to reflect the new capability
. reengineered Fields Management to be more flexible
thus no longer requiring a normal sized tty
Buglets fixed
. forced View_SCROLL off when in 'Batch' mode
. for narrow xterm, fixed return val in show_pmt
. for narrow xterm, fixed cursor pos in linein
. added a window specific end-of-line termcap ptr
( so clr_eol does not overwrite last printable char )
Cosmetic changes
. eliminated TTYGETENVYES define, incorporated as -w
. eliminated RESIZE_LIMIT define, no longer needed
. eliminated FIELD_CURSOR define, no longer possible
. eliminated extra text on Fields Management screen
. fixed miscellaneous man document typos/formatting
Signed-off-by: Jan Görig <jgorig@redhat.com>
Library Changes
. added PROC_EDITCMDLCVT flag
. added an internal (static) fill_cmdline_cvt function:
- reads and "escapes" /proc/#/cmdline
- returns result as a single string in a single vector
- callers are guaranteed a cmdline (no more NULL)
. added vectorize_this_str function, exploited by
fill_cgroup_cvt, fill_cmdline_cvt
. generalized read_cmdline function as read_unvectored, now
exploited by fill_cgroup_cvt, fill_cmdline_cvt, read_cmdline
( cgroup and cmdline no longer need be converted to string )
( vectors before being transformed to final representation )
. fixed bug regarding skipped group numbers (when enabled)
. escape_str made responsible for all single byte translation
with distinction between control chars + other unprintable
. added escaped_copy function for already escaped strings
. reorganized parts of proc_t to restore formatting standards
( displacement changes shouldn't matter with new version # )
. former ZAP_SUSEONLY #define now OOMEM_ENABLE
. added to library.map: escaped_copy; read_cmdline
Top Program Changes
. exploited the new PROC_EDITCMDLCVT provision
. eliminated now obsolete #include "proc/escape.h"
. changed the P_WCH display format if no kernel symbol table
. fixed very old bug in lflgs for out-of-view sort fields
. former ZAP_SUSEONLY #define now OOMEM_ENABLE
Ps Program Changes
. exploited the new PROC_EDITCMDLCVT provision
. exploited the new escaped_copy function
. consolidated pr_args and pr_comm into pr_argcom
Signed-off-by: Jan Görig <jgorig@redhat.com>
This source patchset addresses the following:
Library Extension (readproc)
. added PROC_EDITCGRPCVT flag
. added an internal (static) fill_cgroup_cvt function:
. reads AND parses /proc/#/cgroup
. returns result as a single string in a single vector
. thus no changes to proc_t structure or free memory logic
Program Enhancements (top)
. removed parse_cgroup logic in favor of libproc
. eliminated cgroup sort recurring overhead
. converted WCHAN field to variable width
. generalized variable width field logic in task_show
. real-time notation under P col more distinctive as 'rt'
Program Cosmetic (top)
. CGROUP now known as CGROUPS (plural)
. moved jan's attribution from top.c to 'Notes' in top.h
. numerous comments tweaked
Document Enhancements (top)
. documented CGROUPS field (required renumbering all fields)
. adapted narratives for the 3 current variable width fields
. expanded real-time scheduling notes, 'RT' now shown as 'rt'
Signed-off-by: Jan Görig <jgorig@redhat.com>
Jim Warner