@User, I don't know if this has been fixed since Magic version 8.3.124 but Magic insists upon saving "changes" to unmodified files in my designs. I know we discussed this a little before where timestamps were being updated but it seems that it's just reordering the text in the

.mag

files. I'm upgrading it now but is this an ongoing issue?

diff --git a/columbus/ip/pswitch/lay/pswitch_via_strip.mag b/columbus/ip/pswitch/lay/pswitch_via_strip.mag
index dd27a45..aa2f9ee 100644
--- a/columbus/ip/pswitch/lay/pswitch_via_strip.mag
+++ b/columbus/ip/pswitch/lay/pswitch_via_strip.mag
@@ -7,14 +7,14 @@ rect -591 980 -565 1972
 rect -592 980 -564 1972
 << metal3 >>
 rect -595 980 -561 1416
-use via_m1_m2  via_m1_m2_0
-array 0 0 26 0 30 32
-timestamp 1615675713
-transform 1 0 -591 0 1 983
-box 0 -3 26 29
 use via_m2_m3  via_m2_m3_0
 array 0 0 28 0 5 73
 timestamp 1615675713
 transform 1 0 -595 0 1 980
 box 0 0 34 71
+use via_m1_m2  via_m1_m2_0
+array 0 0 26 0 30 32
+timestamp 1615675713
+transform 1 0 -591 0 1 983
+box 0 -3 26 29
 << end >>

I'll see if I can track this down.

👍🏻

Cell instances may be written in any order because they come from a hash table and so the order is arbitrary; that's not an indication of the cell being modified. There's a flag in each cell that indicates the cell has been modified. In nearly every case, the cell modification is accompanied by a timestamp modification. In short, I can't find any place in the code that would rewrite a file yet maintain all the same timestamps in the output. I need some kind of reproducible example. . .

Ah great. glad you found the cause!

That was a deep rabbit hole. I found not "the cause", but three completely independent causes, all of which had to be tracked down and eliminated separately. But that was well worth the trouble even if it is past my bedtime now.

Haha. Well I'm glad it was worth it. Thanks, @Tim Edwards!