Closed Bug 14929 Opened 27 years ago Closed 26 years ago

colspan cell's should honour percentage width's

Categories

(Core :: Layout: Tables, defect, P5)

x86
Windows 95
defect

Tracking

()

VERIFIED FIXED

People

(Reporter: alan-lists, Assigned: karnaze)

References

()

Details

(Keywords: testcase, Whiteboard: [TESTCASE][nsbeta3+][PDTP5])

Attachments

(4 files)

I am not sure if this is really a bug or just a difference in layout that is not important..... At http://discussion.seattlelab.com/ the center colum that lists discussion groups for SLmail, SLnet, ect is much farther to the right in Mozilla then Netscape 4.61 (Win32) I played with Window sizes ect to make sure nothing there was the problem....
Assignee: troy → karnaze
Component: Layout → HTMLTables
Yes, it does look different. Chris, appears to be a table issue
Whiteboard: [TESTCASE]
Attached file testcase —
Status: NEW → ASSIGNED
I'm testing this with 1999101908 mozilla.exe _and_ viewer.exe on Win98 (btw, I'm testing mats' testcase.) This looks like an incremental-reflow bug visible both in viewer and mozilla. If you resize the window to the left cell gets a litle smaller, even though the table width didn't change. The direction of the resize doesn't matter! If you continue with resizing, the left cell gets smaller until it reaches the right position.
Target Milestone: M15
Bulk moving [testcase] code to new testcase keyword. Sorry for the spam!
Keywords: testcase
I don't see this problem on Win98 2000021108. Can you still reproduce this?
This does not appear to be fixed in today's (21108) win32 build. I can reproduce. It does not break in the same place as 4.x
mass move to m16
Target Milestone: M15 → M16
Moving to M17.
Target Milestone: M16 → M17
Keywords: nsbeta3
seems like a very basic table sizing issue that should be addressed before FCS. could effect many pages.
Whiteboard: [TESTCASE] → [TESTCASE][nsbeta3+]
Target Milestone: M17 → M18
Approving for beta3
If you fill content of cell 2, it will push cell 1 back to 20% (attached testcase). I'm not sure if this is a dup, but there is some similar behavior to bug #19961. In that case, instead of a long image in the first row, the table width is valued at 100% - the similarity is that the TD specific width is not being obeyed.
QA Contact: petersen → chrisd
Priority: P3 → P2
I'm not sure what exactly the bug is here. It is not well defined. Summary says text spacing is farther to the right in Mozilla. I don't know what 'text spacing' refers to but the problem is that the width of the first cell is not being honored. If this is so, then the problem seems to be a dup of #19961. ChrisK, could you take a look and see if you think these bugs are dups? If so, please mark it so. I believe both should be nsbeta3+ bugs if they are not dups. Thanks
I dont think that this bug is dupe of bug 19961. The main story at 19961 is: if you have td with a specified width in px and a second td without any specification and the table is larger than the minimum required width the excess space is distributed also to the cell the with the specified width, making this cell larger than desired. In the last testcase both td have a proportional width specified.
Attached file testcase revisited —
http://bugzilla.mozilla.org/showattachment.cgi?attach_id=13971 shows that this bug occures only with a big image,and does not occure with pure text.
Attached file testcase without image —
Table dump for testcase without white spaces in the colspan: AssignPrelimColWidths en max=9600 count=0 ***START TABLE DUMP*** mColWidths=0 0 col frame cache -> 0=00BA34C4 1=00BA352C **START COL DUMP** colIndex=0 isAnonymous=0 constraint=0 widths=-1 -1 -1 -1 -1 -1 -1 -1 -1 **END COL DUMP** **START COL DUMP** colIndex=1 isAnonymous=0 constraint=0 widths=-1 -1 -1 -1 -1 -1 -1 -1 -1 **END COL DUMP** ***END TABLE DUMP*** AssignPrelimColWidths ex ***START TABLE DUMP*** mColWidths=4950 3795 col frame cache -> 0=00BA34C4 1=00BA352C **START COL DUMP** colIndex=0 isAnonymous=0 constraint=0 widths=900 900 -1 4950 -1 -1 -1 -1 -1 **END COL DUMP** **START COL DUMP** colIndex=1 isAnonymous=0 constraint=0 widths=690 690 -1 3795 -1 -1 -1 -1 -1 **END COL DUMP** ***END TABLE DUMP*** the same if whitespaces are included: AssignPrelimColWidths en max=9600 count=0 ***START TABLE DUMP*** mColWidths=0 0 col frame cache -> 0=02775570 1=027755D8 **START COL DUMP** colIndex=0 isAnonymous=0 constraint=0 widths=-1 -1 -1 -1 -1 -1 -1 -1 -1 **END COL DUMP** **START COL DUMP** colIndex=1 isAnonymous=0 constraint=0 widths=-1 -1 -1 -1 -1 -1 -1 -1 -1 **END COL DUMP** ***END TABLE DUMP*** AssignPrelimColWidths ex ***START TABLE DUMP*** mColWidths=900 690 col frame cache -> 0=02775570 1=027755D8 **START COL DUMP** colIndex=0 isAnonymous=0 constraint=0 widths=900 900 -1 -1 5070 -1 -1 -1 -1 **END COL DUMP** **START COL DUMP** colIndex=1 isAnonymous=0 constraint=0 widths=690 690 -1 -1 3885 -1 -1 -1 -1 **END COL DUMP** ***END TABLE DUMP*** So it seems that the error is inside AssignPrelimColWidths in BasicTableLayoutStrategy.cpp
My understanding of this bug is the following: The big picture or text without whitespaces makes the minimum width of colspanned cell bigger than the minimum required width of both cell's in the second row. This additional space is distributed to both cells based on the ratio of there minimum required width's, but does not take into account that both cell's have a percentage width specified. The in the testcase with the smaller picture or the white space delimited text that mechanism is not invoked as the minimum width of the colspanning cell is smaller than the sum of the minimum width's of both cells in the second row. Disabling this broadening mechanism by simply replacing: http://lxr.mozilla.org/seamonkey/source/layout/html/table/src/BasicTableLayoutSt rategy.cpp#606 if (availWidth > 0) { with: if(PR_FALSE) { Gives perfect rendering of the testcases. This not a patch :-D. It seems that the general problem is in the type mixing, the colspanning cell is auto while the both spanned cell's are percentage cells. Changing summary
Summary: Text spaceing farther to right in Mozilla vs Netscape? → colspan cell's should honour percentage width's
PDT marking P5 because chrisd says Nav has the same problem
Priority: P2 → P5
Whiteboard: [TESTCASE][nsbeta3+] → [TESTCASE][nsbeta3+][PDTP5]
Fixed. Nav does handles all of the test cases. I agree with Buster, that this is fundamental and important. Maybe ChrisD got this mixed up with another bug.
Status: ASSIGNED → RESOLVED
Closed: 26 years ago
Resolution: --- → FIXED
Using 9/12 build, verified fixed.
Status: RESOLVED → VERIFIED
Blocks: 29058
No longer blocks: 29058
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: