https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81106
Bug ID: 81106
Summary: gnatchop -r misplaces BOM
Product: gcc
Version: 8.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: ada
Assignee: unassigned at gcc dot gnu.org
Reporter: simon at pushface dot org
Target Milestone: ---
Host: x86_64-apple-darwin16
Created attachment 41558
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=41558&action=edit
Demonstrator (with BOM)
If gnatchop is given an input file with a UTF8 BOM, it copies the BOM
to the front of each output file.
If the -r switch is given (generate Source_Reference pragmas
referencing original source file), the generated Source_Reference
pragma for the first output file only is output before the BOM, thus
rendering the first output file's BOM ineffective.
I think this is because in gnatchop.adb (Write_Chopped_Files), the BOM
of the first output unit is copied over as part of the beginning of
the input file:
Write_Unit
(Source => Buffer,
Num => Unit_Number,
TS_Time => TS_Time,
Write_BOM => BOM_Present and then Unit_Number /= 1,
Success => Success);
The attachment (zipped because it contains a BOM), if chopped without -r,
contains
$ od -ctx1 pack1.ads
0000000 357 273 277 p a c k a g e P a c k 1
ef bb bf 70 61 63 6b 61 67 65 20 50 61 63 6b 31
0000020 \n w i t h E l a b o r a t e _
0a 77 69 74 68 20 45 6c 61 62 6f 72 61 74 65 5f
0000040 B o d y \n i s \n e n d P a c k
42 6f 64 79 0a 69 73 0a 65 6e 64 20 50 61 63 6b
0000060 1 ; \n
31 3b 0a
whereas with -r it contains
0000000 p r a g m a S o u r c e _ R e
70 72 61 67 6d 61 20 53 6f 75 72 63 65 5f 52 65
0000020 f e r e n c e ( 0 0 0 0 0 1 ,
66 65 72 65 6e 63 65 20 28 30 30 30 30 30 31 2c
0000040 " p a c k 1 . a d a " ) ; \n 273
20 22 70 61 63 6b 31 2e 61 64 61 22 29 3b 0a ef
0000060 273 277 p a c k a g e P a c k 1 \n
bb bf 70 61 63 6b 61 67 65 20 50 61 63 6b 31 0a
0000100 w i t h E l a b o r a t e _ B
77 69 74 68 20 45 6c 61 62 6f 72 61 74 65 5f 42
0000120 o d y \n i s \n e n d P a c k 1
6f 64 79 0a 69 73 0a 65 6e 64 20 50 61 63 6b 31
0000140 ; \n
3b 0a
0000142
The BOM (ef bb bf) starts at 057.
