This morning I had a little disagreement on pshat in Rashi when he says how to measure the bow shaped city that has 4000 amos or more between then ends. I understood that it meant that each person gets 4000 amos from his house so if a guy in middle of the bow leaves from the end, he may only get a few amos. Basically it's as though each house is treated as it's own city so you only square off each house since the city can't be combined.
Others understood it to mean that it depends where you're counting from - basically this is an exception where we won't have a square shaped city. You'll count two thousand from the center and then 2000 from each end. I found that hard to believe that was what Rashi meant. It turned out that our disagreement was actually a machlokes between the Rashba and the Ritva in how to explain Rashi.
Kollel Iyun Hadaf summarizes all the opinions:
(a) RASHI (DH Pachos; 61a, DH Ir ha'Asuyah) writes that "each person measures the Techum from the entrance of his house." Rashi implies that one may walk only 2000 Amos from his house, even though he is still within 2000 Amos of the border of the city. Rashi understands that we ignore the city limits entirely when the city is shaped like a broad bow and we view the house as though it is in an uninhabited desert. (This is the understanding of the RASHBA and MAGID MISHNEH (Hilchos Shabbos 28:8) in Rashi's words.)
(b) The RITVA and RASHBA disagree with Rashi. They point out that in every case of a city, the entire area of the city is considered like one's four Amos with regard to the Shabbos Techum, and the 2000-Amah Techum is measured from the borders of the city. Similarly, for the people living in the bow-shaped city, when the Techum cannot be measured from the "bowstring" it is measured from the borders of the "bow" part of the city (i.e., from the place along the "bow" that is parallel to the person's house). The RITVA asserts that this is also the intention of Rashi -- we measure each person's Techum from the edge of the bow. For those living further inside the bow, we measure 2000 Amos from the point along the bow that is closest to them. (According to this understanding, Rashi mentions "each person's house" only to negate the following two opinions.)
(c) The ROSH cites the MAHARAM MI'ROTENBURG who maintains that the people in the bow may walk 2000 Amos in the direction of the bowstring, starting from the place where the two separate sides of the bow have not yet begun to spread apart from each other, and an imaginary bowstring of less than 4000 Amos can be drawn between the two ends of the bow.
(d) The TUR (OC 398) explains that those in the bow may walk 2000 Amos in the direction of the bowstring starting from a point 2000 Amos away from the bow.
(e) The Rashba quotes the RA'AVAD who points out that one who walks in the direction opposite that of the bowstring may walk not only 2000 Amos from the border of the city, but also 2000 Amos from an imaginary square that is circumscribed around the bow's bend of the city. (That is, this bow-shaped city is measured the same way the Techum of any city, whose borders are not straight lines, is measured. A square is circumscribed around the city, from which the Techum is then measured; see previous Insight.)
HALACHAH: The SHULCHAN ARUCH (OC 398:4) writes only that the Techum is measured from the "bow" part of the city ((b) above). The REMA adds that it is measured from where the bow widens to more than 4000 Amos ((c) above).